Re: forward masking (was: Re: Hugos)

From: Barbara Lamar (altamiratexas@earthlink.net)
Date: Sun Dec 17 2000 - 23:25:07 MST


At 03:19 PM 12/18/2000 +1100, Damien Broderick wrote:
>At 08:07 PM 17/12/00 -0800, J. R. wrote:
>
> >I don't think AIs would mimic our slips, because the glitch (in this case)
> >is in the interface (keyboard), and for AI to make these syntactical errors
> >would require replicating (in an AI android) the inherent deficiencies in
> >human data entry capabilities.
>
>Not so. The slip was in my brain - I knew I was about to write a word
>ending with -ing, and mistakenly modified the one I was currently
>generating to conform with that pattern. I don't *think* the problem is a
>brain-hand lag; I suspect it's inevitable contagion in any sequential or
>look-ahead syntactic process, which an AI would share, I'd imagine.

What Damien writes makes sense. As he said earlier, these same sorts of
errors occur in speech, and I believe similar problems can also occur in
puzzle and problem-solving. On the brighter side, human language has quite
a bit of redundancy built into it. Any English speaker would have known
what he meant when he wrote "being doing" and would have mentally corrected
the error. Surely the AI would also have built-in redundancy,
"double-entry" systems and so forth to catch errors.

Barbara



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:32:26 MST